California Department of Transportation(Caltrans)

Transcription

1 California Department of Transportation(Caltrans) Probabilistic Cost Estimating using Crystal Ball Software "You cannot exactly predict an uncertain future" Presented By: Jack Young California Department of Transportation Division of Engineering Services, Structure Cost Estimates

2 Caltrans Overview Staff 0f 22,000-24,000 Employees Headquarters, 12 Regional Districts and Division of Engineering Services(Structures) 50,000 miles of freeway lanes Supports Intercity Rail Service Many miles of bicycle lanes Maintains and Inspects 24,000 State and Local Bridges Nearly $10 Billion Dollars Under Contract in Construction Helpful Links:

3 Caltrans Organization Chart Structure Cost Estimates Branch Jack Young Senior Bridge Engineer

4 Population of California 44 Million People

5 Cost Estimation Approaches Single Point Estimate Plus Contingency History indicates a high likelihood that the final Project Costs will exceed the estimated amount. To account for possible cost overruns, contingency amounts are added to the total estimate amount. Unfortunately, there is no "probability" associated with the likelihood of meeting the programmed cost! Probabilistic Cost Estimate Reveals the full range of cost and associated probability of possible outcomes.

6 Why Traditional Spreadsheet Analysis Often Fails Traditional spreadsheet analysis uses a single value (like the average) to represent uncertain or variable inputs. The results are static or deterministic and most likely will be unrepresentative of the range of possible outcomes. For example, if it takes you an average of 1.5 hours to get to the airport and thru security and you leave 1.5 hours before your flight takes off, you will miss your plane ~ 50% of the time.

7 Probabilistic Cost Estimate "Challenges and Opportunities" Cost Estimators recognize that uncertainty and risk will occur. Cost estimates are not static or deterministic. They are forecasts that have a range of possible outcomes. Understanding the implications of these ranges leads to more accurate decisions. Can Estimators develop a standard process for risk analysis that identifies "challenges and opportunities" and also quantifies risk in unit pricing?

8 Probabilistic Cost Estimate What is a Model? A model is a combination of data and logic constructed to predict the behavior and performance of a process. Crystal Ball software works with spreadsheet models, specifically Microsoft Excel. A model is a spreadsheet that has taken the leap from being a data organizer to a predictive analysis tool.

9 Probabilistic Cost Estimate "What is Risk?" Uncertainty about a situation can often indicate risk, which is the possibility of loss, damage, or any other undesirable event. Most Project Managers and Estimators desire low risk, which would translate to a high probability of success and cost savings.

10 Probabilistic Cost Estimate "What is Uncertainty?" Uncertainty is assessed in cost estimate models for the purpose of estimating the risk (probability) that a specific funding level may or may not be exceeded. Points to keep in mind when analyzing risk: How likely is the risk? How significant is the risk? Where is the risk? How do I manage the risk?

11 A Better Way To Analyze Uncertainty Simulation with Crystal Ball Use ranges as inputs Thousands of outcomes with associated certainty Easy to analyze and communicate Range of Outcomes What-if Analysis Even increments of values Multiple outcomes but no associated probabilities Difficult to analyze and very time consuming Range Estimates Most-likely, best-case, worst-case 3 outcomes Single point estimate One representative value as input 1 outcome Value of Analysis (Betterness)

12 Probabilistic Cost Estimate "Simulation" What is Simulation? The application of models to predict future outcomes with known and uncertain inputs. Why use Simulation? Measure the behavior of the outcomes given changes in the inputs. Simulation can be considered a probabilistic framework for analysis

13 Inputs: What is Monte Carlo simulation? A computer simulation of N trials calculating multiple (hundreds, thousands) scenarios of your model by repeatedly sampling random combinations of your uncertain inputs. Outputs: Sampling statistics characterize output variation (mean, standard deviation, etc.) Predict and Quantify ranges of output(probability/forecast of cost) Identify primary variation drivers (sensitivity analysis)

14 How Does Crystal Ball Appear in MS Excel? Toolbar Define Menu Run Menu Analyze Menu Setup simulation Views results and creates reports Runs the simulation

15 Which Assumption Curves Do I Use? Use distributions based on past historical data or physical principles (normal, log-normal) More realistic, Less conservative Use expert opinion to develop triangular distribution (Minimum, Most Likely, Maximum value) Use bounds with uniform distribution (Minimum, Maximum value) Less realistic, More conservative

16 Structure Cost Estimate - Crystal Ball Example Apply or develop probability curves to a historical data set Build a model identifying potential threats and opportunities Simulate project bidding 10,000 times using Monte Carlo Simulation Develop a probabilistic estimate forecast including a range of likely final costs with associated confidence levels Conduct sensitivity analysis and determine largest contributors to cost variance

17 Overview - Crystal Ball Cost Estimate Model (Input and Output)

18 MODEL INPUT

19 ASSUMPTION CURVE - INPUT

20 FORECAST CHART - OUTPUT

21 FORECAST PERCENTILES - OUTPUT

22 SENSITIVITY CHART - OUTPUT

23 In Summary: Probabilistic Cost Estimate - Pros: Forecast results reveal the full range of cost and probability of possible outcomes. Sensitivity analysis indicates which inputs drive most of the output variation (threats and opportunities). Forces the Estimator to refine the historical item cost data. Employs "quantitative" risk management strategies in unit pricing. Transparent and defensible input and report output results. Excellent backup for cost estimate certification. Premium reporting and charting features for Executive summaries. Potential to reduce cost estimate contingencies. User friendly software(excel Overlay).

24 Crystal Ball Software Uses in Transportation Project Cost Estimating FHWA-Cost Estimate Reviews: Cost Estimate Reviews (Mega projects) Marin-Sonoma Narrows, I-5 HOV, I-10 HOV, I-215 HOV, I-405 HOV, Caldecott 4 th Bore Tunnel Project, Doyle Drive, etc. California Department of Transportation Division of Engineering Services Structures Development of all phases of structure cost estimates Depositions Transparent/Defensible Back-Up for Cost Estimate Certification QC/QA/IQA Practices

25